Characterizing habitat preference of Eurasian river otter (Lutra lutra) in streams using a self-organizing map

Abstract

We studied the habitat preferences of Eurasian river otters (Lutra lutra) using the distribution patterns of the numbers of spraints and sprainting spots of otters, as well as related environmental variables (habitat zone, river management, bank type, vegetation coverage, width, depth, etc.) in two streams. The numbers of otter spraints and sprainting spots were sampled monthly in two streams on Geoje Island, Republic of Korea, from January to December 2004. Additional environmental variables were measured at the sampling sites. A self-organizing map (SOM), which is an unsupervised artificial neural network, was used to characterize the habitat preferences of otters. In our results, the SOM classified three different groups of study sites based on their habitat conditions, and the habitat differences were effectively visualized on the trained SOM map. Otters showed spatial and temporal dynamics in the numbers of spraints and sprainting spots, and revealed habitat preferences for shallow, narrow areas of streams and edges of water that were not far from reservoirs but covered with trees and shrubs. Additionally, otters preferred an environment in which weirs reduced the drift of water and gathered fishes and had a natural type of stream bank; these findings are relevant for river management. Otters adapted to places close to roads, residential areas, and agricultural areas with some tolerance of human interference.